Despite the fact that hemorrhage from esophagogastric varices in patients with alcoholic liver disease is a major health problem in the United States today, the most effective and economical method of managing this life threatening complicating has not been determined. Therefore, the main objective (specific aim 1) of this proposal is to determine in a randomized controlled trial which of two commonly employed therapeutic modalities, portasystemic shunting (surgical) or endoscopic variceal sclerosis (non-operative) is more effective in the management of alcoholic cirrhotics who bleed from varices. Of specific interest will be the relative influences of these two treatment options on length of survival, quality of survival, psychoneurological function, and medical costs. Three shunting operations, distal splenorenal, end-to-side portacaval, and side-to-side portacaval shunts, will be used in the portasystemic shunt group, and selection of procedure will be based on the patient's hemodynamic status as judged from preoperative visceral angiography. Secondary objectives of this investigation are to determine the comparative effects of portasystemic shunt operations and endoscopic variceal sclerosis on the following variables: specific aim 2, hepatic portal perfusion, total hepatic blood flow, and portal pressure; specific aim 3, intestinal absorption and psychoneurological function and the relationship between these two variables; specific aim 4, hepatic functional reserve as measured by three quantitative tests of hepatocellular function.